Napari plugin to ease manual correction of epithelia segmentation in movies
To analyse individual cell trajectory from epithelia movies marked for cell-cell junctions, a very precise segmentation and tracking is required. Several tools such as TissuAnalyzer, Epyseg, CellPose or Dist2Net perform very good segmentation (~5% of errors). However, this still implies a high amount of cells to correct manually.
EpiCure allows to decrease the burden of this task. Several features are proposed to ease the manual correction of the segmented movies, such as error detection, numerous shortcuts for editing the segmentation, option for tracking, display and measure/export options. EpiCure detect segmentation errors by taking advantage of temporal information. When a correction is done at a given frame, EpiCure relink the track to adjust for the changes.
See the full documentation here
To install EpiCure on a fresh python virtual environment, type inside the environement:
pip install epicure
Then launch Napari, and the plugin should be visible in the Plugins list.
If you already have an environment with Napari installed, you can also install it directly in Napari>Plugins>Install/Uninstall plugins
To have the code to be able to modify it, clone this repository. You can use pip install -e . so that everytime you update the code, the plugin will be updated.
The input files of EpiCure can be already tracked or not. Tracking options are proposed in EpiCure:
- Laptrack centroids
- Laptrack overlaps
Refer to the documentation for documentation of the different steps possible in the pipeline.
If you use EpiCure, thank you for citing our work: Gaëlle Letort, Léo Valon, Arthur Michaut, Tom Cumming, Laura Xénard, Minh-Son Phan, Nicolas Dray, Curtis T. Rueden, François Schweisguth, Jérôme Gros, Laure Bally-Cuif, Jean-Yves Tinevez, Romain Levayer; EpiCure (Epithelial Curation): a versatile and handy tool for curation of epithelial segmentation. Development 2026; dev.205701. doi: https://doi.org/10.1242/dev.205701
The dataset presented in our main EpiCure publication is available on zenodo: 10.5281/zenodo.20607705.
EpiCure is mainly developed in CNRS UMR3738, in the Developmental and Stem Cell Biology Department of Institut Pasteur. If you have a question on using EpiCure or ask to add a feature, either file an issue or write in the imagesc forum.
Any contribution is most welcome. Do not hesitate to contact us beforehand through filing an issue (choose "Other questions/comments" type). To suggest the addition of a new feature, you can also contact us by filing an issue choosing the "Feature request" option.
If you encounter a code related issue using EpiCure, please file an issue in this repository.
To facilitate contributions, we added a pixi file in the repository. Developers can run:
pixi run buildto install locally the plugin and test it manuallypixi run testto run locally the tests from thesrc/testsdirectorypixi run build-docto locally build the documentation pages. The documentation will be build in thesitefolder in the repository. The filesite\index.htmlcan be open in a browser to see the generated documentation.
The code API can be found here in the online documentation. Examples of using EpiCure without opening the interface can be found in the test files and the repository notebooks.